Evidence for involvement of the voltage-dependent Na+ channel gating in depolarization-induced activation of G-proteins.

Evidence for activation of pertussis-toxin-sensitive G-proteins by membrane depolarization in rat brainstem synaptoneurosomes was recently reported (Cohen-Armon, M., and Sokolovsky, M. (1991) J. Biol. Chem. 266, 2595-2605; (1991) Neurosci. Lett. 126, 87-90) and is further supported in this study by the observation that the depolarization-induced effect is inhibited when G-proteins are stabilized in the non-activated state with guanosine 5'-O-(2-thiodiphosphate) (GDP beta S), which was introduced into synaptoneurosomes during the process of permeabilization and resealing. In the present study, agents that either keep the voltage-dependent Na+ channel in persistently activated state (while Na+ currents are blocked) or prevent it from activation were used in an attempt to determine whether the voltage-dependent Na+ channels are involved in the depolarization-induced activation of pertussis-toxin-sensitive G-proteins. The main probe employed was the cardiotonic and antiarrhythmic agent DPI, which is a racemic mixture of two enantiomers, one of which (the R enantiomer) reportedly prevents depolarization-induced activation of the Na+ channel while the other (the S enantiomer) inhibits Na+ channel inactivation. The results suggest that while inactivation of the voltage-dependent Na+ channel does not interfere with the putative depolarization-induced activation of G-proteins, membrane depolarization affects G-proteins and the coupled muscarinic receptors only if the voltage-dependent Na+ channels are capable of being activated. Thus, inhibition of the depolarization-induced activation of Na+ channels was accompanied by inhibition of the depolarization-induced activation of pertussis-toxin-sensitive G-proteins and by modifications of both the coupling of G-proteins to muscarinic receptors and the ADP-ribosylation of Go-proteins. These effects could be counteracted by persistent activation of the voltage-dependent Na+ channels (while Na+ current was blocked). Our observations may suggest that the voltage-dependent Na+ channel gating is involved in the depolarization-induced activation of pertussis toxin-sensitive G-proteins and may provide evidence for a possible mechanism of membrane depolarization signal transduction in excitable cells.

Evidence for activation of pertussis-toxin-sensitive G-proteins by membrane depolarization in rat brainstem synaptoneurosomes was recently reported (Cohen-Armon, M., and Sokolovsky, M. (1991) J. Biol. Chem. 266, 2595-2605; (1991) Neurosci. Lett. 126,[87][88][89][90] and is further supported in this study by the observation that the depolarization-induced effect is inhibited when G-proteins are stabilized in the nonactivated state with guanosine 5'-0-(2-thiodiphosphate) (GDPBS), which was introduced into synaptoneurosomes during the process of permeabilization and resealing. In the present study, agents that either keep the voltage-dependent Na+ channel in persistently activated state (while Na+ currents are blocked) or prevent it from activation were used in an attempt to determine whether the voltage-dependent Na+ channels are involved in the depolarization-induced activation of pertussis-toxin-sensitive G-proteins. The main probe employed was the cardiotonic and antiarrhythmic agent DPI, which is a racemic mixture of two enantiomers, one of which (the R enantiomer) reportedly prevents depolarization-induced activation of the Na+ channel while the other (the S enantiomer) inhibits Na+ channel inactivation. The results suggest that while inactivation of the voltage-dependent Na+ channel does not interfere with the putative depolarizationinduced activation of G-proteins, membrane depolarization affects G-proteins and the coupled muscarinic receptors only if the voltage-dependent Na+ channels are capable of being activated. Thus, inhibition of the depolarization-induced activation of Na+ channels was accompanied by inhibition of the depolarization-induced activation of pertussis-toxin-sensitive G-proteins and by modifications of both the coupling of Gproteins to muscarinic receptors and the ADP-ribosylation of Go-proteins. These effects could be counteracted by persistent activation of the voltage-dependent Na+ channels (while Na+ current was blocked). Our observations may suggest that the voltage-dependent Na+ channel gating is involved in the depolarizationinduced activation of pertussis toxin-sensitive G-proteins and may provide evidence for a possible mechanism of membrane depolarization signal transduction in excitable cells. We have recently reported evidence for a depolarizationinduced activation of PTX'-sensitive G-proteins (hereafter termed G-proteins) in brain-stem synaptoneurosomes. These observations revealed the same features as those observed in activation of G-proteins by neurotransmitters and hormones (1)(2)(3)(4). They included, first, a depolarization-induced enhancement of the specific binding of [3H]GTP/[3H]GDP to membranes of permeabilized-resealed synaptoneurosomes (5), attributed to the exchange of GDP bound to inactivated Gproteins for labeled GTP (6)(7)(8)(9). The enhancement in GTP binding could be prevented by PTX-catalyzed ADP-ribosylation of G-proteins and reversed by repolarization of the depolarized synaptoneurosomes (5). Second, we observed a depolarization-induced conversion of the muscarinic receptor from a high affinity to a low affinity state for agonist binding (5,10). This effect was attributed to a depolarization-induced activation of G-proteins (5, lo), since it could be imitated by persistent activation of the G-proteins coupled to the muscarinic receptor (11,12) and prevented by distortion of the coupling between receptor and G-proteins following PTXcatalyzed ADP-ribosylation of G-proteins (5,10). Third, we observed a depolarization-induced inhibition of PTX-catalyzed ADP-ribosylation of the a-subunit of Go-proteins (13). Since ADP-ribosylation of PTX-sensitive G-proteins can be inhibited by dissociation of the a-from the @?-subunits (14,15), an event which occurs on G-protein activation (1)(2)(3)(4), the observed inhibition of ADP-ribosylation on membrane depolarization may also indicate a depolarization-induced activation of these G-proteins (13). As previously reported, all of these depolarization-induced effects on G-proteins could be reversed by membrane repolarization, were independent of the release of transmitters (5, 13), and were not dependent on the response of serotoninergic, a-and @-adrenergic, or NMDA receptors (5).
Additional evidence in support of a depolarization-induced activation of PTX-sensitive G-proteins comes from the observation in this study that GDPOS, when introduced into synaptoneurosomes (5) in order to prevent G-protein activation (3, [6][7][8][9], prevents the depolarization-induced effect on the muscarinic high affinity agonist binding and on PTXcatalyzed ADP-ribosylation of G-proteins. In the present work we investigated the possible involvement of the voltage-dependent Na' channel (hereafter referred to as the Na' channel) in the observed depolarizationinduced activation of G-proteins. This was done by the use of agents known to affect the Na+ channel gating, including the inhibitor of Na+ channel inactivation, batrachotoxin (BTX) (16)(17)(18), the local anesthetic tetracaine (19 and Refs. therein), and the cardiotonic and antiarrhythmic drug 4-[3-(4-diphenylmethyl-l-l-piperazinyl)-2-hydroxy-propoxy]-1H-indole-2carbonitrile (DPI-201-106) (Sandoz), a racemic mixture of the enantiomers DPI-205-429 ( R enantiomer) and DPI-205-430 ( S enantiomer) (20). Determination of the influence of DPI on the depolarization-induced effect on G-proteins was of great interest, since the R enantiomer prevents the depolarization-induced activation of the Na' channel, while the S enantiomer, by inhibiting Na+ channel inactivation, prolongs its activated state (20,21). The enantiomers compete in an allosteric manner with one another and with labeled BTX for binding to the Na' channel (20,21). The effects of the two enantiomers of DPI on the depolarization-induced activation of G-proteins were examined by their effect on the high affinity binding of [3H]AcCh to the muscarinic receptor and the specific binding of [3H]GTP/[3H]GDP to membranes of permeabilized-resealed synaptoneurosomes, on the basis of our previous results (5, 13).
Our results indicate that inactivation of Na+ channels, induced by prolonged membrane depolarization (22), does not interfere with the putative depolarization-induced activation of G-proteins. On the other hand, prevention of the depolarization-induced activation of the Na+ channels (i.e. the channel remains in the closed state (22)) prevents the depolarization-induced activation of PTX-sensitive G-proteins, modifies the coupling of G-proteins to the muscarinic receptors, and inhibits PTX-catalyzed ADP-ribosylation of the a-subunit of Go-proteins. These events can be counteracted by agents that induce persistent activation of the Na+ channel (with Na' current blocked by TTX) but do not themselves induce G-protein activation.
It should be noted that according to current findings, the inactivated state of the Na+ channel is considered not to be identical with the "closed state" of the Na+ channel although in both states Na+ current is zero (17,22 and Refs. therein).
The findings of the present study support involvement of the voltage-dependent Na+ channel gating in the depolarization-induced activation of PTX-sensitive G-proteins. The results of this study are also in line with previous indications ( 5 , 10) that unmodified ADP-ribosylation sites must be present on the a-subunit of PTX-sensitive G-proteins in order for the effect of membrane depolarization on G-proteins to be induced. The apparent involvement of both Na+ channel gating and ADP-ribosylation sites on PTX-sensitive G-proteins in the putative depolarization-induced activation of Gproteins may provide evidence for a possible mechanism of membrane depolarization signal transduction in excitable cells.  (33.6 Ci/mmol) were purchased from Du Pont-New England Nuclear. D(-)-2-amino-5-pbosphovaleric acid was from Cambridge Research Biochemicals (U.K.). Succinic acid and EDTA were from Merck (Darmstadt, Germany). Dowex 2 X 10, 50-100 mesh (Cl-form) was from Fluka AG (Chemische Fabrik, Buchs, Germany). Antibodies against the N terminus of Go-protein (GC/2) were purchased from Du Pont.

Reagents
Brain Tissue Preparation-Adult male rats of the CD strain were obtained from Levinstein's Farm, Yokneam, Israel, and maintained as described previously (23). From pooled brain-stem regions obtained from 3-to 4-month-old rats (IO), synaptoneurosomes were prepared according to Hollingsworth et al. (24) in Krebs-Henseleit buffer containing 118.5 mM NaC1, 4.7 mM KCI, 1.18 mM MgCIz, 2.5 mM CaCI2, 24.9 mM NaHC03, 10 mM glucose, and 1.18 mM KHzPOl in an atmosphere of 95% 02, 5% COz at 25 "C. The synaptoneurosomes were then resuspended in the same buffer in the absence of CaCI2. Under the experimental conditions, synaptoneurosomes retained their membrane potential for up to 4 h (5).
PHIAcCh Binding Assay-Binding of [3H]AcCh to the muscarinic receptor in the synaptoneurosomal membrane preparation was measured according to the method of Gurwitz et al. (25), with a number of modifications (5). Ca2+-free Krebs-Henseleit buffer in which CaC12 was replaced by NaCl ([Ca'"] <20 p~) was used (5, 26). Following incubation for 20 min with the cholinesterase inhibitor diisopropylfluorophosphate (300 p~) , synaptoneurosomes (20-4 aliquots) were added to tubes containing 20 pl of buffer and the indicated concentration of [3H]AcCh. After 1 h of incubation with gentle shaking at 25 "C, the reaction was terminated by the addition of 5-6 ml of icecold buffer and filtration under high pressure through GF/C filters (Whatman, 25-mm diameter). The filters were immediately washed with 6 ml of buffer and then counted for tritium using a scintillation mixture (Hydroluma) and a scintillation spectrometer (LKB 1218) at 48% efficiency. Nonspecific binding was determined in the presence of 1 p~ atropine. All determinations were carried out in triplicate and were found to vary by no more than 15%.
The measured binding of [3H]AcCh to the muscarinic receptor in this preparation probably does not include [3H]AcCh uptake or internalization, since it was found to be similar in synaptoneurosomal preparations and in membranes prepared from synaptoneurosomes either by homogenization or by lysis in hypotonic buffered solution containing 50 mM Tris-HC1 and 1.8 mM MgCl,, pH 7.4.
Under the experimental conditions employed, binding of [3H]AcCh to the nictonic receptors did not occur, as indicated by measurement of the specific binding of [3H)AcCh in the presence and absence of the specific nicotinic blocker di-tybocurarine (27). Since di-tybocurarine did not exert any effect on either the specific or the nonspecific binding of [3H]AcCh measured in the presence of 1 p~ atropine (25, 28), it was concluded that [3H]AcCh binding measured under these experimental conditions represented the specific binding of [3H]AcCh to the muscarinic receptor only.
PHINMPB Binding-Binding of the muscarinic antagonist [3H] NMPB to the muscarinic receptor in synaptoneurosomes was assayed as previously described (28). Nonspecific binding was determined in the presence of 1 p~ atropine (23).
Depolarization Experiments-Synaptoneurosomes were depolarized by exposure to high concentrations of the selectively permeable cation, K+, in the Ca2+-free Krebs-Henseleit buffer solution through an exchange of Na+ for K+. In most experiments membrane depolarization was achieved by an increase in [K+] from 4.7 to 50 mM.
Washings and filtrations were carried out in cold high [K+] buffers.
In our previous reports (5, 10, 13), we described experiments designed to test the possibility that high [K'] has a direct effect on [3H]AcCh binding, [3H]GTP binding, and PTX-catalyzed ADP-ribosylation. These experiments included parallel measurements of the above three parameters, in synaptoneurosomes and in membrane preparations, in the presence of various concentrations of K' and Na+. The effects of high [K'] observed in the synaptoneurosomes were not observed in membrane preparations, suggesting that the effects observed in synaptoneurosomes are attributable to high [K+induced membrane depolarization rather than to a direct effect of high [K+] on the binding of [3H]AcCh or [3H]GTP/[3H]GDP or on ADP-ribosylation of G-proteins (5,13).
A possible effect of inward Na+ current induced by the high [K+]induced depolarization was ruled out, since voltage-dependent Na+ channels are inactivated during the prolonged membrane depolarization (22) induced by exposure to high [K'], and since the potent Na+ current blocker TTX (22) did not interfere with the depolarization-induced effects on G-proteins and muscarinic receptors (5, 10).
Estimation of Membrane Potential by PHJTPP Accumuhtion- [3H]TPP permeates freely across cell membranes, and thus its distribution at equilibrium in various systems is dependent on the membrane potential (29,30). Changes in membrane potential were confirmed by monitoring of the intracellular accumulation of 13H]TPP' according to the method of Cheng et a2. (31). Synaptoneurosomes (approximately 4 mg of protein/ml) were incubated with approximately 7 X 10.' M [3H]TPP+ (40 pl) at 25 "C for 20 min. The method is described in detail elsewhere (5).
Treatment of Synaptoneurosomes with BTX and TTX-Synaptoneurosomes were preincubated in Ca2+-free Krebs-Henseleit buffer with BTX (1 pM) for 40 min at 37 "C. In order to exclude Na+ currents, TTX (1 WM) was added together with BTX (10,32,33). The reaction was terminated by centrifugation (1000 X g, 10 min) and two washes with the same Ca*+-free buffer. Following this pretreatment, no changes in membrane potential (as measured by [3H]TPP+ accumulation) or uptake of "Na+ into the pretreated synaptoneurosomes were observed (5, lo), i.e. inward Na+ current was blocked and hence membrane depolarization was not induced under these experimental conditions (22).
Binding of [JHJGTPIrHJGDP or GDPPS to Permeabilized-resealed Synaptoneurosomes-Synaptoneurosomes were permeabilized in the presence of ATP as described in detail previously (5). The procedure included incubation of the synaptoneurosomes with ATP (6 mM) in isotonic buffer containing 169 mM glucose, 40 mM NaC1,4.7 mM KC1, 1.18 mM KH2P04, 24.9 mM NaHC03, 0.5 mM succinate, pH adjusted to 8.3-8.4 (permeabilization buffer). Permeabilization was carried out in the presence of 0.2 p~ [3H]GTP (5). After precisely 40 min at 25 "C, the synaptoneurosomes were resealed by two successive cycles of incubation (10 min) with Krebs-Henseleit buffer, and centrifugation (1000 X g, 5 min) (5). They were then exposed for 1-2 min to Krebs-Henseleit buffer containing either 4.7 or 50 mM [K+]. In order to minimize binding of the labeled guanine nucleotide to the external surface of the synaptoneurosomes, samples (100-111 aliquots, approximately 400 pg of protein) were loaded onto an anion exchanger (DOWEX 2 X 10, 50-100 mesh, quaternary ammonium type of structure, C1-form, equilibrated with Krebs-Henseleit buffer containing the desired [K']). The samples were eluted with 1 ml of Krebs-Henseleit buffer containing either 4.7 or 50 mM [K+] and filtered under pressure through GF/C filters (Whatmanj. Loading of labeled GTP/GDP was then measured. Binding of 13H]GDP to the synaptoneurosomal membranes was measured following lysis of the synaptoneurosomes in distilled water during filtration. The method was described in detail previously ( 5 ) . In some experiments, synaptoneurosomes were permeabilized in described above. The synaptoneurosomes were then subjected to [3H] AcCh binding experiments and ADP-ribosylation according to the procedures described.
Binding of PHJCTP to Membrane Preparations-Membranes were prepared from synaptoneurosomes by homogenization in hypotonic buffer (50 mM Tris-HC1, pH 7.4), with the use of a motor-driven Teflon pestle (950 revolutions/minute) in a glass homogenizer. The membrane preparations were incubated with [3H]GTP, at concentrations ranging from 0.1 to 2.2 p~, in Krebs-Henseleit buffer containing <20 p~ Ca2', at 25 "C for 20 min (equilibrium was reached after 10 min). Each sample (100 pl) contained approximately 400 pg of protein.
Incubation was terminated by the addition of ice-cold buffer. Samples were filtered through GF/C filters and counted for tritium in a presence of 300 PM Gpp(NH)p. The method was described in detail scintillation spectrometer. Nonspecific binding was measured in the previously (5). PTX Treatment of Synaptoneurosomes-Synaptoneurosomes in Ca2+-free Krebs-Henseleit buffer were treated with 200 ng/ml PTX (2 h, 37 "C, 95% 02,5% Con). The PTX-treated synaptoneurosomes were then washed in the same buffer and subjected to various experiments, as described below. Similar treatment with PTX was applied to permeabilized-resealed synaptoneurosomes that were either loaded with GDPPS (200 p~) in the permeabilization buffer or unloaded.
Immunoblotting-Gels were electroblotted onto nitrocellulose paper overnight at 10 "C and a constant current of 150 mA, as described by Towbin et al. (37). Dried nitrocellulose strips were immersed first in 3% gelatin and then in anti-a, GC/2 (38) (1:500 dilution) in 5% bovine serum albumin, as described elsewhere (39). The bands with bound antibodies were visualized by a peroxidase-conjugated second antibody (37).

Effects of Membrane Depolarization on the High Affinity
Binding of pH/AcCh to Permeabilized-resealed Synaptoneurosomes Loaded with GDPPS-In order to assess the effect of membrane depolarization on G-protein activation (5, 13), we examined the depolarization-induced effect on the conversion of agonist-binding sites in the muscarinic receptor from high to low affinity (5) in the presence of GDPBS, which keeps Gproteins in an unactivated state (3, 6-9). Because the cell membrane is not permeable to GDPPS (40), the latter was introduced into synaptoneurosomes which were permeabilized and resealed for the purpose, as described in "Materials and Methods." Effective perrneabilization and resealing was confirmed by 13HJTPP+ accumulation measurements, as described previously (5). Specific binding of [3H]AcCh to the resealed synaptoneurosomes, either unloaded or loaded with GDPPS, was measured in buffers containing 4.7 or 50 mM [K+], i e . at resting or depolarization membrane potential, respectively. Changes in membrane potential were confirmed by measurement of [3H]TPP+ accumulation (see "Materials and Methods"). The effect of membrane depolarization on the high affinity binding of [3H]AcCh to muscarinic receptors in unloaded permeabilized-resealed synaptoneurosomes is presented in Fig. 1A. Fig. IB shows the corresponding effect in synaptoneurosomes loaded with GDPPS. As demonstrated, membrane depolarization reduced the high affinity binding of ['HIAcCh in the unloaded synaptoneurosomes (Fig. LA). This effect could be reversed by re-exposure of the depolarized synaptoneurosomes to 4.7 mM [K+] buffer (not shown). Similar results were previously observed in intact synaptoneurosomes (5, 10) and were attributed to a depolarization-induced conversion of high affinity muscarinic agonist-binding sites to a low affinity state, presumably reflecting activation of Gproteins coupled to the muscarinic receptors (5,7,12). In synaptoneurosomes loaded with GDPpS, the high to low affinity conversion did not occur (Fig. lB), thus further supporting the involvement of G-protein activation in the observed depolarization-induced effect.
As reported previously, under these experimental conditions no measurable release of AcCh was detected (5,101. The Effect of Membrane Depolarization on PTX-catalyzed ADPribosylation of Permeabilized-resealed Synaptoneurosomes Loaded with GDPPS-The effect of membrane depolarization on PTX-catalyzed ADP-ribosylation in brain-stem synaptoneurosomes was previously reported (13). In that study, membranes prepared from synaptoneurosomes that were ADPribosylated in high [K+] buffer underwent intense [32P]ADPribosylation upon subsequent exposure to PTX-catalyzed ADP-ribosylation in the presence of [32P]NAD, suggesting that membrane depolarization had inhibited the initial ADPribosylation possibly as a result of a depolarization-induced activation of PTX-sensitive G-proteins (14,15). This possibility was further supported by examination of the depolarization-induced effect on ADP-ribosylation in the presence of GDPPS (3, 6-9). Brain-stem synaptoneurosomes were permeabilized ( 5 ) , loaded with GDPpS (200 p~ in the permeabilization buffer), and resealed, as described under "Materials and Methods." GDPPS-loaded and unloaded permeabilizedreleased synaptoneurosomes were ADP-ribosylated by PTX (200 ng/ml, 37 "C, 2 h, 95% 02, 5% C 0 2 ) . Membranesprepared from loaded or unloaded synaptoneurosomes were then subjected to an additional ADP-ribosylation, performed in the presence of PTX A-protomer and [32P]NAD (see "Materials and Methods"). membrane depolarization. Similar inhibition of ADP-ribosylation was previously observed in depolarized nonpermeabilized synaptoneurosomes (13). The fact that loading of the synaptoneurosomes with GDPBS counteracted a depolarization-induced inhibition of the ADP-ribosylation of 39-kDa Gproteins suggests that such inhibition involves a depolarization-induced activation of the PTX-sensitive G-proteins.

Effects of DPI Enantiomers on Depolarization-induced Changes in the Binding of PHJAcCh to the Muscarinic Recep-
tor-Involvement of the Na' channel gating in the observed depolarization-induced modulation of high affinity agonist binding (5) was examined in the presence of each of the enantiomers of DPI. Synaptoneurosomes prepared from rat brain-stem were depolarized by exposure to high-[K'] buffer (50 mM), as described under "Materials and Methods." Changes in membrane potential were monitored by measurement of [3H]TPP+ accumulation (5). As previously observed (5), high affinity binding of [3H]AcCh (25) to control synaptoneurosomes a t resting potential was decreased upon membrane depolarization by high [K+]; this effect could be reversed by re-exposure of the synaptoneurosomes to 4.7 mM [K'] buffer (resting potential). This depolarization-induced effect was stereospecifically affected by the Na' channel effector DPI (20, 21) (see Fig. 3, A-C). Fig. 3A demonstrates the effect of the DPI S enantiomer (5 p~) , in the presence of T T X (1 p~) , and the effect of T T X (1 p~) on the depolarization-induced decrease in binding of [3H]AcCh to brain-stem synaptoneurosomes. As shown, the S enantiomer, which inhibits inactivation of the Na' channel or prolongs its activated state (20,21). did not affect the depolarization-induced conversion of binding sites in the muscarinic receptor from high to low affinity (Fig. 3A). Since Na' current was blocked by T T X (32, 33). induction of membrane depolarization by inward Na' currents was prevented (22). T T X (1 p~) did not affect the depolarization-induced high to low affinity conversion, as also observed previously (10). However, in the presence of the R enantiomer (5 p~) , which prevents activation of the Na' channel (20,21). the depolarization-induced high to low affinity conversion was completely prevented (Fig. 3H ), and treatment with the racemic mixture of the drug (5 p~) partially inhibited the depolarization-induced affinity conversion (Fig. 3C).
Since the DPI agents did not affect the total binding to the muscarinic receptors, measured by the binding of the antagonist ['HINMPB (not shown), these results appear to provide evidence for the possible involvement of Ne' channel activation in the depolarization-induced conversion of muscarinic high affinity sites into the low affinity state.

Effects of DPI on Binding of PHJAcCh to the Muscarinic Receptor in Membrane Preparations-In order to distinguish
the effects of DPI enantiomers on the binding of ['HIAcCh to the muscarinic receptor from their effects on the depolarization-induced changes in ['HIAcCh binding, binding of ('HI AcCh in the presence of the enantiomers was measured in membrane preparations. Treatment with the R and with the S enantiomer of DPI had similar effects on the density of high affinity binding sites of ['HH]AcCh in membranes prepared from rat brain-stem (Fig. 4). Binding of ['HIAcCh to brain-stem membranes treated with the S and R enantiomers of DPI is presented in Fig. 4, A and C, respectively. The corresponding Scatchard plots are presented in the imets.
The calculated changes in ['HIAcCh-binding site density (B,,,.J and in the apparent dissociation constant ( K d ) induced by the S and R enantiomer are presented in Fig. 4, R and D, respectively ( n = 10). As shown, in the presence of either the S or the R enantiomer up to a concentration of 10 p M , the decrease in density of high affinity binding sites ( R*,,,.x/R,,,.x) was negligible. Binding affinity, however, increased in the presence of the S enantiomer ( Fig. 4 A , inset, and R ) . An effect similar to that was observed after treatment with BTX (28) (which also inhibits inactivation of the Na' channel (16)(17)(18)). No significant change in ['HIAcCh binding affinity was observed in the presence of the R enantiomer (Fig. 4R, inset, and D 1. In our subsequent experiments aimed a t determining the effects of DPI on the depolarization-induced changes in ['HI AcCh binding, we used concentrations a t which the binding density of ['HIAcCh to the muscarinic receptor in membrane preparations is not affected. These concentrations are similar to those used to induce prolonged activation ( S enantiomer) or to prevent activation of the Na' channel ( R enantiomer) in electrophysiological studies (20,21).   Fig. 5 shows the effects of the R enantiomer on the depolarization-induced decrease in high affinity binding of [3H]AcCh to the muscarinic receptor in control and in BTX-pretreated preparations. As shown, BTX prevented the inhibition exerted by the R enantiomer on the depolarization-induced decrease of [3H] AcCh high affinity binding (see also Fig. 3B).

Effect of the R enantiomer of DPI on the Depolorizationinduced Affinity Changes in the Muscarinic Receptor after Pretreatment with
It should be noted that BTX in the presence of TTX had no effect on the depolarization-induced conversion of muscarinic high affinity sites to the low affinity state (Fig. 5,  inset). Since inward Na+ current (and hence additional membrane depolarization) was excluded under these experimental conditions, as confirmed by measurements of [3H]TPP accumulation and '*Na+ uptake by the pretreated synaptoneurosomes (not shown) (see Ref. lo), exposure of BTX-and TTXpretreated synaptoneurosomes to high [K+] buffers induced membrane depolarization, as it did in control preparations. As in our previous study (lo), we could therefore observe the depolarization-induced effect on [3H]AcCh high affinity binding when the Na+ channels were in a persistently activated state (Fig. 5 , inset).

1[3H] AcCh] ( n M )
The fact that both the DPI S enantiomer (Fig. 3C) and BTX pretreatment counteracted the inhibition imposed by the DPI R enantiomer on depolarization-induced high to low affinity conversion (Figs. 3B and 5) may suggest that the depolarization-induced conversion of muscarinic agonistbinding sites from high to low affinity could be affected by changes induced in the gating state of the voltage-dependent Na+ channels, i.e. in the inactivated state the Na+ channels do not appear to interfere with the depolarization-induced affinity conversion. Neither inactivation induced by the exposure to high [K+] (prolonged depolarization (22)), nor inhibition of Na+ channel inactivation by BTX (16)(17)(18) and DPI S enantiomer (20,21) interfered with the depolarizationinduced affinity conversion ( Fig. 3A and Fig. 5 , inset). However, when the depolarization-induced activation of the Na+ channels was prevented by the DPI R enantiomer (20, 21), the depolarization-induced high to low affinity conversion was blocked (Fig. 5 ) . The fact that this inhibition could be counteracted by persistent activation of the Na' channels ( Fig. 5 ) may suggest that Na+ channel activation is involved in the depolarization-induced muscarinic high to low affinity conversion.
Effect of D P I on the Depolarization-induced Binding of Labeled GTPIGDP to Permeabilized-resealed Synaptoneurosones-The depolarization-induced increase previously ob-   (Fig. 6, upper panel) and intrasynaptoneurosomal binding (Fig. 6, lower panel) were measured after exposure (1-2 min) of the resealed synaptoneurosomes to buffers containing either 4.7 or 50 mM [K'] in the absence (control) and presence of the racemic mixture of DPI and each of its enantiomers (see "Materials and Methods"). Changes in membrane potential were confirmed by measurements of [3H]TPP+ accumulation ( 5 ) . As shown, similar amounts of [3H]GTP/[3H]GDP were loaded into the various preparations (Fig. 6, upper panel). In control preparations, the amount of [3H]GTP/[3H]GDP binding to depolarized synaptoneurosomes (in 50 mM [K' ] buffer) was approximately 4-fold higher than the binding to synaptoneurosomes at resting potential (in 4.7 @M [K+] buffer) (Fig. 6,

lower panel). A similar depolarization-induced enhancement in [3H]GTP/[3H]GDP binding was previously reported ( 5 ) .
The effects of the racemic mixture and enantiomers of DPI on the depolarization-induced increase in intrasynaptoneurosomal binding of labeled GTP/GDP are presented in Fig. 6  (lower panel). In the presence of the R enantiomer ( 5 ~L M ) ( i e . where activation of Na' channels was prevented (20, 21)), no depolarization-induced increase in [3H]GTP/[3H]GTP binding was detected. On the other hand, the S enantiomer ( 5 pM), which was added in the presence of TTX (1 p~) to exclude inward Na' current (10, 32, 33), did not affect the depolarization-induced enhancement in [3H]GTP/[3H]GDP binding which was observed in control preparations (Fig. 6,  lower panel). Under these experimental conditions, TTX by itself did not interfere with [3H]GTP/[3H]GDP binding (not shown). In the presence of the racemic mixture ( 5 p~) the depolarization-induced enhancement in [3H]GTP/[3H]GDP binding was partially inhibited (Fig. 6, lower panel).
These results may suggest the involvement both of ADPribosylation sites on PTX-sensitive G-proteins and of the voltage-dependent NaC channel gating in the depolarizationinduced enhancement of [3H]GTP/[3H]GDP binding attributed to G-protein activation ( 5 ) .
These measurements were all carried out in the absence of   [6][7][8][9] in the presence and absence of DPI, is presented in Fig.  7. The calculated apparent dissociation constant (Fig. 7, inset) for [3H]GDP binding to control membranes was 0. amounts of [3H]GTP bound/milligram of protein were previously measured in this preparation ( 5 ) . As shown, the R or S enantiomer of DPI did not significantly alter either the specific or the nonspecific binding of 13H]GDP, i.e. neither enantiomer affected the displacement of GDP by 13HJGTP in the membrane preparations (Fig. 7). Also pretreatment with 1 PM BTX, which resulted in activation of Na' channels in these membrane preparation (10, 281, did not affect the binding of [3H]GTP/[3H]GDP to the membranes (not shown). These results may indicate that neither activation nor inactivation of the voltage-dependent Na' channel (22) by itself induces activation of G-proteins as reflected by the displacement of GDP by [3H]GTP (1)(2)(3)(4)(6)(7)(8)(9).
Effects of DPI on Coupling of the Muscarinic Receptor to Gproteins-Uncoupling of the muscarinic receptor from Gproteins, triggered by G-protein activation (2,41,42), induces conversion of the muscarinic receptor from a high affinity to a low affinity state for agonist binding (11, 12). Persistent activation of G-proteins should therefore be accompanied by a low affinity state of [3H]AcCh binding. In order to determine whether agents that affect the gating state of the Na' channel also affect the coupling between the muscarinic receptor and G-proteins, we examined whether persistent activation of Gproteins can affect the muscarinic affinity for [3H]AcCh (11,12,25) in the presence of the DPI R enantiomer (i.e. under conditions where Na' channel activation is prevented) or S enantiomer (i.e. when Na' channels are persistently activated) (20,21). Fig. 8A shows the Gpp(NH)p-induced decrease in high affinity binding of [3H]AcCh in control membrane preparations. Since no parallel change was observed in the density of muscarinic binding sites, as measured by the binding of the antagonist [3H]NMPB (not shown), the decrease in [3H]AcCh binding was attributed to a conversion of agonist-binding sites in the muscarinic receptor from a high to a low affinity (10, 28). Similar results were obtained in the presence of the DPI racemic mixture ( 5 1 M ) (Fig. 8B) or the S enantiomer (5 p M ) (not shown). In the presence of the R enantiomer ( 5 p M ) , however, no such Gpp(NH)p-induced decrease in high affinity binding of [3H]AcCh was observed (Fig. 8 B ) . It thus appears that in the presence of the R enantiomer the coupling between muscarinic receptors and G-proteins was distorted, so that the muscarinic receptor remained in its high affinity state even when the G-proteins were persistently activated in the presence of Gpp(NH)p.

Effects of Modifiers of the Voltage-dependent Na' Channel on p2P]PTX-catalyzed ADP-ribosylation of G-proteins-The
ADP-ribosylation site on cyo-and ai-proteins apparently plays an essential role in the coupling between G-proteins and muscarinic receptors (41)(42)(43)(44)(45). Therefore, in view of the observed effect of the R enantiomer on the coupling of the muscarinic receptors to G-proteins (Fig. 8 ) , we examined the effects of agents capable of modifying the gating state of the Na' channel on PTX-catalyzed ADP-ribosylation of G-proteins.
Persistent activation of the Na' channels, without induction of Na' current, was obtained by pretreatment of synaptoneurosomes with BTX or the DPI S enantiomer in the presence of T T X (32, 33). Inhibition of Na' channel activation was induced by the DPI R enantiomer (20,21) or the local anesthetic tetracaine (19,46). Membranes prepared from synaptoneurosomes incubated with these agents (37 "C, 1 6) while Na+ current was blocked. Weak 32P-labeling was detected in a 30 kDa protein band. The effects of the above modifiers of the Na+ channel gating on the labeling of this protein band were similar to those observed for the 39-kDa G-proteins (Fig. 9A). Fig. 9B, lanes 1-4, illustrate the inhibitory effect of the potent local anesthetic drug tetracaine (19,46) on [32P]ADP-ribosylation of the 39-kDa G-proteins in membranes prepared from synaptoneurosomes incubated with increasing concentrations of tetracaine (0.5, 1, 2, and 5 p h i ) . The effective concentrations of tetracaine were similar to those obtained for blocking of Na' channels and for inhibition of the binding of labeled BTX to the Na' channel in brain preparations (46,47). The effects induced by tetracaine on the 32P-labeling of a 30-kDa protein band were similar to those observed for the 39-kDa G-proteins (Fig 9B).
[32P]ADP-ribosylation of membranes prepared from control and PTX-pretreated (100 ng/ ml, 37 "C, 2 h, 95% 02, 5% COz) synaptoneurosomes is presented in lanes 5 and 6, respectively. As expected for PTXsensitive G-proteins, labeling of the 39-kDa protein band was weak in membranes prepared from ADP-ribosylated synaptoneurosomes, thus indicating ADP-ribosylation of these proteins in PTX-treated synaptoneurosomes. 32P-Labeling of the 30-kDa protein band was, however, not affected. Fig. 9C compares [32P]ADP-ribosylation performed in synaptoneurosomal membranes in the presence and absence of the reducing agent DTT (20 mM). It should be noted that [32P]ADP-ribosylation of the membranes was performed in the presence of the PTX A-protomer, so DTT was not required for activation of the ADP-ribosyltransferase subunit of PTX (35,45). Fig. 9C presents the autoradiograms of ["PI ADP-ribosylated membranes following SDS-PAGE analysis when ADP-ribosylation was conducted in the presence of 20 mM DTT (lune 1 ) and in its absence (lanes 2-4). Like the 39-

Immunolabeling of f2P]ADP-ribosylated G-proteins in Membranes Prepared from Synuptoneurosomes Pretreated with Modifiers of the Voltage-dependent Na' Channel-Im-
munolabeling was performed in order to identify the 39-kDa G-proteins modified by agents affecting the Na' channel gating. Synaptoneurosomes were pretreated with the Na' channel activators BTX (16)(17)(18) and the S enantiomer of DPI (20, 21) (Na' current was blocked by the addition of 1 p~ TTX), or with the inhibitors of Na+ channel activation, i.e. the R enantiomer of DPI (20,21) or tetracaine (19, 46).
Membranes prepared from control and pretreated synaptoneurosomes were subjected to [32P]ADP-ribosylation in the presence of the A-protomer of PTX and then analyzed by SDS-PAGE (10% polyacrylamide). The resulting blots ("Western" blots (37)) were exposed to autoradiography, then immunostained with antibodies to the N terminus of the asubunit of Go-proteins. These antibodies were specifically chosen in order to avoid possible interference by the ADPribosylation of a,-proteins assumed to be on the C terminus (39) with an immunological reaction at the C terminus of the a,-subunit. Labeling of this protein band was previously observed following reaction with antibodies to the C terminus of the a-subunit of Go-proteins (5, 10, 13). Fig. 1OA presents autoradiograms of the [D2P]ADP-ribosylated 39-kDa G-protein in control membranes (prepared from untreated synaptoneurosomes) (lane 1 ) and in membranes prepared from synaptoneurosomes treated with T T X (1 p~) and either the S enantiomer of DPI (5 p~) (lane 2) or BTX (1 p~) (lane 3). from synaptoneurosomes treated with tetracaine (1 p~) following pretreatment with BTX (1 p~) and T T X (1 p~) (lane 4 ) , and from synaptoneurosomes pretreated with the R enantiomer (5 p~) (lane 6). In lane .5, membranes prepared from untreated synaptoneurosomes were subjected to ["PIADP-ribosylation in the absence of DTT. As shown, no [32P]ADP-ribosylation was achieved in this non-reducing environment. Fig. 10R presents immunoblots of the same preparations following the reaction with antibodies to the N terminus of the a-subunit of Go-proteins. As shown, all preparations reacted similarly with the Go-antibodies. Thus, the various modifications of the Na+ channel gating apparently did not affect the amounts of a,-subunit in the membrane, while specifically affecting their ability to be [32P]ADP-ribosylated in the presence of PTX A-protomer. As shown, Go-proteins in membranes prepared from synaptoneurosomes subjected to a persistent opening of Na+ channels with Na+ current prevented (lane 3) were more efficiently [32P]ADP-ribosylated than the rest. This effect was partially inhibited following incubation of the synaptoneurosomes with tracaine (1 p~) , which inhibits BTX binding to the Na' channel (46,47) (lane 4 ) . Thus, in membranes prepared from synaptoneurosomes pretreated with inhibitors of the Na+ channel activators (lanes 4 and 6 ) the [32P]ADPribosylation of a,-proteins was inhibited.
It should be noted that although the 39-kDa G-protein was equally immunolabeled in the presence (lanes 1-4 and 6 ) and absence of DTT (lane 5), no [32P]ADP-ribosylation was achieved in the non-reducing environment, suggesting that cleavage of S-S bonds in the a,-protein is essential for PTXcatalyzed ADP-ribosylation to occur.

DISCUSSION
The results presented in this study provide evidence for involvement of the voltage-dependent Na+ channel gating in depolarization-induced activation of PTX-sensitive G-proteins. The effects of membrane depolarization on G-proteins and on muscarinic receptors were previously reported in brain-stem synaptoneurosomes and myocytes (5,13). Inhibition of the depolarization-induced effect on G-proteins by GDPBS (Figs. 1 and 2), which causes persistent inactivation of G-proteins (1)(2)(3)(4). supports the contention that membrane depolarization induces G-protein activation. The present results demonstrate that this putative depolarization-induced activation of G-proteins can be inhibited by agents that prevent the activation of the voltage-dependent Na+ channel (Figs. 3, 5, and 6). It should be noted that only the channel gating was modified here, as induction of Na' current and its consequent changes in membrane potential were prevented by addition of the potent specific Na+ channel blocker TTX (1 p~) (32,33). Prevention of the voltage-dependent Na+ channel activation was accompanied by distortion of the coupling of G-proteins to the muscarinic receptor (Fig. 8) and inhibition of the PTX-catalyzed ADP-ribosylation of Goproteins ( Figs. 9 and 10). Accordingly, agents that inhibit the Na' channel inactivation (22), such as BTX (16)(17)(18) and the S enantiomer of DPI (20, 21) (with Na' current abolished), could counteract the inhibition induced by inhibitors of the Na' channel activation such as the DPI R enantiomer (20,21) and the local anesthetic tetracaine (46) (Figs. 3C, 5,6,8R, 9, A and B, and 10).
Since prolonged membrane depolarization induced by exposure of the synaptoneurosomes to high [K'] buffer result9 in inactivation of the voltage-dependent Na' channel (22). the depolarization-induced effect on G-proteins cannot be attributed to inward Na' currents. Moreover, it appears that inactivation of the Na' channels does not interfere with the depolarization-induced activation of G-proteins (Figs. 1-3 and 5-7 and Refs. 5,10,13). However, prevention of activation of the Na' channel (22) did abolish the depolarization-induced effects on G-proteins and on muscarinic receptors (Figs. 38, 5,6,8B, 9, A and B, and 10).
The fact that agents that prevent activation of the Na' channel appear to modify both the coupling of G-proteins to muscarinic receptors (Fig. 8) and the ADP-ribosylation of Goproteins (Figs. 9, A and B, and 10). and the evidence that PTX-catalyzed ADP-ribosylation also prevents depolarization-induced activation of G-proteins, may suggest that the ADP-ribosylation sites on PTX-sensitive G-proteins are involved in the depolarization-induced activation of G-proteins.
The fact that BTX and DPI-201 affect Na' channels in a reversible manner (16-21) does not necessarily exclude the possibility that the effects on ADP-ribosylation of G,-protein, which were detected in membranes prepared from pretreated synaptoneurosomes, were mediated by persistently activated or persistently blocked Na' channels. One cannot exclude a possible secondary effect induced by reversible changes in the Na' channel and causing irreversible changes in the ADPribosylation sites; interference with an endogenous ADPribosylation of G-proteins would be reflected in the subsequent ["PIADP-ribosylation performed in membranes pre-pared from the pretreated synaptoneurosomes. However, because of the hydrophobicity of these effectors, it is possible that they are not removed by washes and homogenization, and this could explain their sustained effect on the [32P]ADPribosylation of membranes prepared from the treated synaptoneurosomes. In a previous study, our attempts to remove [3H]BTX from brain homogenates met with difficulties (10). Similar inhibition of ADP-ribosylation of Go-proteins in the presence of tricyclic antidepressants was recently reported (48).
Previous reports of inhibition by local anesthetics of the formation of second messengers, such as adrenergically induced CAMP formation (49) and PI turnover (50,51), are consistent with the present results and may suggest that the effect of these agents on the formation of second messengers is mediated by their modification of the coupling of G-proteins to receptors (and possibly also to other effectors in the membrane), which may be achieved in turn by modification of the ADP-ribosylation site on G-proteins.
The compatibility between the effects induced by membrane depolarization on the muscarinic receptors and on the coupled G-proteins (Figs. 1 and 2, 3 and 6, 5 and 9, and 10) may support previous evidence (5) for an effect of membrane depolarization on G-proteins in the postsynaptic membrane, which houses the muscarinic receptors (11, 12). The possibility that other receptors coupled to G-proteins are also changed in response to membrane depolarization should be investigated. Such a phenomenon might be of major significance with regard to signal transduction in the central nervous system.
The suggested modifications of the ADP-ribosylation site(s) on the a-subunit of Go-proteins, induced by changes in the state of the Na+ channel gating, are consistent with previously reported evidence suggesting a mutually positive effect of the activation of the Na' channel and agonist high affinity binding to the muscarinic receptor mediated by G-proteins (10, 28,52,53). Thus, net TTX-blockable and atropine-blockable 22Na+ influx was induced by carbamylcholine in brain-stem synaptoneurosomes, when the muscarinic receptor was in the high affinity state (IO), i.e. coupled to G-proteins (2,41,42,44). Also, the binding of labeled BTX to activated Na+ channels (16)(17)(18) in membranes prepared from rat brain-stem and atria was increased in the presence of muscarinic agonists (10, 28). Reciprocally, activation of Na+ channels by pretreatment with BTX increased the binding affinity of [3H]AcCh for the high affinity sites of the muscarinic receptor (28,52) in these preparations. These mutual effects could all be abolished by persistent activation of G-proteins or by uncoupling of the muscarinic receptor from the G-proteins, i,e. in the low affinity state of the muscarinic receptor (10,28). It should be noted that PTX-catalyzed ADP-ribosylation of the synaptoneurosomes (200 ng/ml PTX, 37 "C, 2 h, 95% 02, 5% CO,) did not abolish this effect and therefore apparently does not induce uncoupling of the muscarinic receptor from G-proteins (5, lo), whereas it did prevent the depolarization-induced exchange of GDP by [3H]GTP on G-proteins ( Fig. 6 and Refs. 5, 10). This may suggest that PTX-sensitive G-proteins contain more than one type of ADP-ribosylation sites, which are involved in the interaction of G-proteins with the muscarinic receptor and in the response to membrane depolarization. The fact that Go-proteins are ADP-ribosylated by PTX-A promoter only in a reducing environment (Figs. 9C and 10) may also suggest that this reaction involves changes in the spatial structure of the a,-subunit and may therefore point to the involvement of several sites on this protein in the ADP-ribosylation. This suggestion is in line with recent observations (54).
A one-to-one stoichiometry in the interference with Gprotein activation induced by inhibition of Na+ channel activation seems unlikely, since the density of Na+ channels measured in the membrane of brain preparation is lower than the density of G-proteins (1-2 pmol/mg protein (16,55) versus 40 pmol/mg protein in this preparation or 100-200 pmol/mg protein in various brain preparations (2,56). However, a similar discrepancy exists also with respect to the density of receptors known to activate G-proteins versus the density of G-proteins in the membrane, i.e. the density of muscarinic receptors in brain preparation is 200 fmol/mg protein (brainstem) and 1 pmol/mg protein (cortex) (25, 28). A possible explanation was recently proposed by Coulter and Rodbell (57), who suggested on the basis of cross-linking performed in membranes of brain synaptoneurosomes that groups of Gproteins ("multimeric structures") might interact with receptors, while a single G-protein in the multimeric complex might be activated by a single receptor at a time. Such multimeric structure of G-proteins allows for signal amplification (57).
Figs. 9 and 10 present labeling of a 30-kDa protein band which was affected similarly to that of the 39-kDa G-protein by pretreatment of the brain-stem synaptoneurosomes with agents that modify the voltage-dependent Na+ channel. Labeling of this 30-kDa protein band in the presence of ["PI NAD could be detected only in a reducing environment (ix. in the presence of 20 mM DTT) and even in the absence of P T X (Fig. 9, B and C). Further studies are now being carried out in an attempt to identify the 30-kDa protein band and find out whether it has any function in a possible coupling between Na+ channels and Go-proteins in brain tissue. The &subunit (30-40 kDa) of the voltage-dependent Na' channel complex found in brains of most vertebrates (58), and which couples to the a-subunit of the Na+ channel complex (260 kDa) (55,58,59) by disulfide bonds (58), is one of the proteins that might be represented by the 30-kDa protein band.
In summary, according to previously reported evidence (5, 13) and the present results, membrane depolarization may induce G-protein activation similarly to activation of G-proteins induced by binding of ligands to G-protein-coupled receptors (1-3, 6-9). The results presented here provide evidence that the Na+ channel gating may be involved in membrane depolarization-induced activation of PTX-sensitive Gproteins in brain-stem synaptoneurosomes. In terms of this model, since the voltage-dependent Na+ channel is the first element known to respond to membrane depolarization under physiological conditions (22,56), membrane depolarization signals that induce a threshold potential (22) in the postsynaptic membrane (60 and Refs. therein) may be transduced into a "message" in the postsynaptic excitable cell as a result of G-protein activation. Elucidation of the molecular mechanism causing the depolarization-induced activation of G-proteins may provide information on a depolarization signal transduction in excitable cells.